I design multi-agent platforms for banks and financial institutions: credit workflows, document intelligence, risk, and the governance that makes any of it deployable. Alongside the day job, I keep a long-running study practice in mathematics, machine learning, philosophy, and the contemplative traditions.
This site is the working library where that study accumulates: articles, research writing, and a reading curriculum. It's maintained for my own use and shared in case it's useful to yours.
On architecture, deployment, the major cloud platforms, and the questions where engineering meets the study of mind.
The research literature ordered for reading, not reference, from Shannon in 1948 through current work on agents, alignment, and interpretability.
Careful maps of the recent literature, combined with a production architecture perspective: governed autonomy, the economics of inference, memory and drift.
The running log of work worth knowing about, year by year. The curriculum is ordered for study; the radar tracks the moving edge.
One Sanskrit word traced across six schools of Indian philosophy, an interactive companion to Lakshmi Kapani's study.
A reader's atlas of John Kreiter's thirteen books, techniques catalogued, cross-referenced, and put in order.
I've spent around fifteen years designing and delivering AI systems for financial institutions across Asia-Pacific and Europe, most recently multi-agent platforms for credit, document understanding, fraud, and regulatory work, where the interesting problems are as much about governance and trust as they are about models.
The study practice runs alongside: mathematics and machine learning from first principles, and a serious interest in how attention and awareness actually work, approached through psychology, philosophy of mind, and the contemplative traditions that have been mapping this territory far longer than we have.
I keep the two threads on one site because, in practice, they inform each other. If we're building systems that reason alongside people, the design constraints aren't only computational.
Designing agentic AI deployments for a large financial institution, orchestration, LLM gateways, governance.
Working through the papers curriculum in order, and writing study guides as I go.
Slowly expanding the articles, architecture and deployment first, the consciousness material as it matures.
Prāṇāyāma, meditation, and somatic work, the empirical side of the consciousness research.